Terence Hines - Pseudoscience and the Paranormal

Impressed with these seemingly powerful and consistent results, Marks and Kammann (1980) attempted to replicate Targ and Puthoff’s remote-viewing results. In a series of thirty-five studies, Marks and Kammann were unable to replicate the results. Their results always showed chance performance. Naturally feeling that they were doing something wrong, they searched for differences in procedure between their studies and those of Targ and Puthoff that might explain the difference in results. The crucial variable turned out to be unexpected. Targ and Puthoff provided their judges with unedited transcripts of the subjects’ impressions, including the comments made by the experimenter who stayed with the subject. Marks and Kammann edited their transcripts to remove cues that would enable a judge to match the subjects’ transcribed comments to the location. Such cues can provide a great deal of information.

Examination of the few actual transcripts published by Targ and Puthoff (1977; Wilhelm 1976) show that just such clues were present. To find out if the unpublished transcripts contained cues, Marks and Kammann wrote to Targ and Puthoff requesting copies. It is almost unheard of for a scientist to refuse to provide his data for independent examination when asked, but Targ and Puthoff consistently refused to allow Marks and Kammann to see copies of the transcripts. Marks and Kammann were, however, able to obtain copies of the transcripts from the judge who used them. The transcripts were found to contain a wealth of cues. In addition, the judge had also been given a list of the locations visited in the order in which they were visited. A simplified example will demonstrate how this information could easily have been used to correctly match the transcripts to the locations, with no paranormal powers needed. Suppose you are a judge in a remote-viewing experiment in which only three locations are visited by the demarcation team. You are given three transcripts, A, B, and C, which are in random order. You are also given—as was the judge in the Targ and Puthoff experiment—a list of the order in which the three target locations were visited: (1) university library, (2) supermarket, and (3) large bridge. Transcript A contains the phrase “Third time’s the charm.” Transcript C contains the phrase “Don’t be nervous—you’re just starting.” You can now correctly assign each transcript to the correct location. Transcript A is the large bridge, B is the supermarket, and C is the library. You’ve achieved an accuracy of 100 percent without even reading any of the actual descriptions in the transcripts. It would be very difficult for a judge to ignore such cues, even if he or she were trying to do so.

Marks and Kammann (1980) showed, in a procedure they wryly called “remote judging,” that subjects were able to match the transcripts to the correct locations using only the cues provided. When these cues were eliminated, but the description of the subjects’ impressions remained intact, matching fell to a chance level.

In 1980 Charles Tart (Tart, Puthoff, and Targ 1980) claimed that a rejudging of now-edited transcripts from one of Targ and Puthoff’s earlier experiments still resulted in above-chance performance. However, Targ and Puthoff again refused to provide copies of the actual transcripts used in this study. They suppressed this vital evidence until July 1985, when it was finally made available. The transcripts still contained numerous cues, and Marks and Scott (1986) concluded that “considering the importance for the remote viewing hypothesis of adequate cue removal, Tart’s failure to perform this basic task seems beyond comprehension. As previously concluded, remote viewing has not been demonstrated in the experiments conducted by Puthoff and Targ, only the repeated failure of the investigators to remove sensory cues” (p. 444).

Another aspect of the remote viewing situation was noted by Marks and Kammann (1980) and deserves mention. Their attempts to replicate the results of Targ and Puthoff were, as noted, complete failures. But before the judging was done, the subjects were given feedback about what target location corresponded to which of their descriptions. They frequently became convinced that their descriptions had been extremely accurate. Marks and Kammann note that descriptions in remote-viewing experiments are vague, mentioning trees, buildings, water, and sky. Given the vagueness of the descriptions, when subjects are told or shown the location, they will always be able to find some points of similarity between the location and the description. Thus, like cold readers becoming convinced that they can divine hidden knowledge from their clients, subjects in a remote viewing experiment become convinced that their descriptions are paranormally inspired.

The remote-viewing controversy lasted more than a decade. It is a sobering example of how sloppy experiments and the conclusions based on them can be accepted as evidence in parapsychology. It further demonstrates the great amount of hard work it takes to put such erroneous conclusions to rest.

The German word ganzfeld means “blank field.” The procedure in a ganzfeld study of psi powers is in some ways quite similar to the procedure in a remote-viewing experiment. Both types of studies use a subject trying to pick up impressions telepathically and a sender trying to send them. In the ganzfeld study, however, the sender is thinking about an object (tree, orange, dollar bill, whatever) rather than a location. The subject is seated in the “blank field” trying to receive the sensory impressions. The blank field may be a large, opaque, white screen covering the subject’s entire field of vision, or it may be simply a table tennis ball, cut in half and fixed over the subject’s eyes. The idea is that the ganzfeld cuts out extraneous sensory inputs and permits the telepathic “message” to be read more clearly by the subject who, at the appointed time, gives his impressions of the object the sender is concentrating on.

The ganzfeld technique was developed in the 1970s, and proponents of psi phenomena contend that it represents a method that gives consistent and repeatable evidence for such phenomena. Sound familiar?

Hyman (1985b) reviewed forty-two studies using the technique (essentially the entire corpus of published studies up to the time of the review) and concluded, “I believe that the ganzfeld psi data base, despite initial impressions, is inadequate either to support the contention of a repeatable study or to demonstrate the reality of psi. Whatever other value these studies may have for the parapsychological community, they have too many weaknesses to serve as the basis for confronting the rest of the scientific community” (p. 38).

Hyman (1985b) found several different flaws in the ganzfeld studies. There was inadequate randomization of the targets in a large percentage of the studies. Opportunities for information on the target to inadvertently reach the subject and breaches in security that might have permitted cheating were also present in many of the studies. Many studies were not described in sufficient detail to allow evaluation of what actually happened in the study. Statistical problems were also common. These included the use of incorrect statistical tests and procedures and “multiple testing” errors in which the subjects’ responses were tested several times against chance, using different criteria for scoring the responses for different tests. This, like testing numerous astrological predictions or numerous predictions of any theory, increases the chance of spuriously obtaining a significant result. Further, the greater the number of flaws in a study, the more likely it was to find a significant effect. This is another example of the phenomenon noted earlier, that evidence for ESP and related phenomena disappears as the tightness of experimental controls is increased. On the basis of Hyman’s review, the ganzfeld studies cannot be said to provide evidence for ESP, as claimed by proponents.